Abstract: Mitral valve replacement (MVR) is a critical surgical intervention for severe valvular dysfunction, yet the hemodynamic consequences of prosthetic valve selection remain incompletely understood. This study investigates the impact of bioprosthetic and bileaflet mechanical valves on left ventricular (LV) flow patterns and pressure dynamics using an in vitro platform replicating physiological conditions. A 3D-printed silicone LV model, integrated with porcine mitral valve (including chordae tendineae and papillary muscles) and aorta, was cyclically actuated via a mock circulatory loop. Four configurations were tested: native valve, bioprosthetic valve, and mechanical valves oriented parallel/vertical to the native annulus. Time-resolved 3D flow fields were captured using tomographic particle image velocimetry (Tomo PIV), while a physics-informed neural network (PINN) framework was employed to reconstruct dynamic pressure fields under moving boundary conditions. Key findings revealed distinct hemodynamic profiles: the bioprosthetic valve generated a coherent diastolic vortex ring aligned with native flow direction, but exhibited elevated transvalvular jet velocity (0.93 m/s vs. 0.62 m/s) . This may impose higher fluid-induced wall stress, potentially triggering adverse remodeling. In contrast, the mechanical valve exhibited diastolic pressure distribution comparable to the native state, but flow reversal occurred due to interactions among triple jets generated by its bileaflet structure, particularly in the vertical orientation where asymmetric vorticity distribution dominated. Both prosthetic types failed to propagate vortices to the apical region, increasing stasis risk. The native valve demonstrated optimal efficiency, with apical vortex penetration minimizing blood residence time. By integrating Tomo PIV with PINN-based pressure reconstruction, this study overcomes limitations of traditional methods in handling deforming boundaries, enabling synchronized 3D flow-pressure analysis post-MVR. Clinically, compared with mechanical valves, bioprosthetic valves better maintain organized flow patterns but require careful monitoring of high shear stress-induced potential damage to blood cells; mechanical valves, while exhibiting superior durability, may exacerbate adverse ventricular remodeling due to complex flow patterns. These insights underscore the need for personalized valve selection based on hemodynamic profiling. Future work will incorporate patient-specific geometries and pathological parameters to refine predictive models and prosthesis design.